Inorg. Chem. 1987, 26, 4140-4142
4140
Contribution from the Pacific Northwest Laboratory, Richland, Washington 99352
Tborium(1V) Hydrous Oxide Solubility Jack L. Ryan* and Dhanpat R a i
Received March 9, 1987
The hydrolysis of thorium(1V) has received a considerable amount of study,'q2 but the agreement between some of the individual workers apparently is poor. Most of this work has been carried o u t a t pH values below about 4.5 although one solubility study3 has been carried out in the pH range 3.4-7.0. From this work,3 solubility product values for hydrous Tho2and hydrolysis constants for formation of Th(OH)3+and Th(OH),O(aq) were r e p ~ r t e d . ' . ~The minimum solubility of active (hydrous) Thoz due t o formation of Th(OH),O(aq) was proposed to be M, and t h e minimum solubility of fully crystalline Thoz (above p H 6 ) has been proposed' to be 10-9.6M. Recently, the formation constant and free energy of formation of t h e Th(OH)5-ion have been estimated4 from reported' values of formation constants for the Th(OH)3+,Th(OH)?+, Th(OH)3+, and Th(OH),O(aq) complexes. On the basis of currently accepted thermochemical data5-' for Thoz and for water and OH- ion, t h e reported4value for Th(OH)< would indicate a totally unreasonable solubility of Th02(c) of 0.4 M at unit OH- activity. Similar amphoteric behavior and formation constant values of U(IV),1343*-11 Np(IV),'oJ' and P u ( I V ) ' ~ ~ - ~ -have '' been proposed for formation of t h e M(OH)5-ion and in the case of U and Pu t h e M(OH)62ions.4 All of the tetravalent actinides can be expected to show qualitatively similar hydrolysis behavior with t h e extent of hydrolysis, as reflected by the hydrolysis constants, increasing with increasing a t o m i c number and decreasing ionic radius. As with t h e radius change,12 the degree of hydrolysis might be expected to change most rapidly with the first members of the series. We have recently measured the solubilities of uranium13 and n e p t u n i ~ min ' ~ alkaline solutions under reducing conditions a n d found no convincing evidence of measurable amphoteric behavior, thereby placing much lower limits on formation constants for anionic species of U(1V) and Np(1V) t h a n those reported. I t is very difficult t o maintain U in the reduced (tetravalent) state under high-pH conditions and somewhat less difficult in t h e case of neptunium. Thorium is exclusively tetravalent, and thus its hydrolytic properties at high pH are easier to determine. This work reports results of a study of the solubility of amorphous, hydrous Thoz over the pH range 3.5-14.2. Experimental Section Reagents. Baker reagent grade Th(N03)4-5H20,Eastman 10% tetraethylammonium hydroxide, and Cerac High Purity grade LiOH were used. The LiOH was analyzed for Si by an ion-coupled plasma spectroscopy method and was found to contain 64 pmol of Si/mol of LiOH. All other chemicals were reagent grade. The carbonate concentrations of the tetraethylammonium hydroxide and a stock solution of the LiOH ( I ) Baes, C. F.; Mesmer, R. E. The Hydrolysis of Cations; Wiley: New York, 1976 (and references therein). (2) Brown, P. L.; Ellis, J.; Sylva, R. N. J . Chem. Soc., Dalton Trans. 1983, 31. (3) Nabivanets, B. I.; Kudritskaya, L. N. Ukr. Khim. Zh. 1964, 30, 891. (4) Barnum, D. W. Inorg. Chem. 1983, 22, 2297. (5) Fuger, J. MTP In?. Reu. Sci. 1972, Series One, 7, 157. (6) Oetting, F. L.; Rand, M. H.; Ackerman, R. J. The Chemical Thermodynamics of Actinide Elements and Compounds-I. The Actinide Metals; International Atomic Energy Agency: Vienna, 1976. (7) Wagman, D. D.; Evans, W. H.; Parker, V. B.; Schumm, R. H.; Halow, I.; Bailey, S. M.; Churney, K. L.; Nuttell, R. L. J . Phys. Chem. Ref. Data 1972, 11, Suppl. No. 2. (8) Langmuir, D. Geochim. Cosmochim. Acfa 1978,42, 547. (9) Lemire, R. J.; Tremaine, P. R. J . Chem. Eng. Data 1980, 25, 361. (10) Allard, B.; Kipatsi, H.; Liljenzin, J. 0.J . Inorg. Nucl. Chem. 1980, 42, 1015. ( 1 1) Phillips, S.L. "Hydrolysis and Formation Constants at 25 OC"; Report LBL-14313; Lawrence Berkeley Laboratory: Berkeley, CA, 1982. (12) Keller, C. The Chemistry of the Transuranium Elements; Verlag Chemie GmbH: Weinheim, FRG, 1971; p 125. (13) Ryan, J. L.; Rai, D. Polyhedron 1983, 2, 947. (14) Rai, D.; Ryan, J. L. Inorg. Chem. 1985, 24, 247.
0020- 1 6 6 9 / 8 7 / 1326-4140$01.50/0
B
Dl
were determined by BaC03 precipitation, and sufficient strontium nitrate was added to the tetraethylammonium stock solution to make it 3 X lo-' M in excess Sr2+and to the LiOH stock solution to make it 8.5 X M in excess Sr2+after S r C 0 3 precipitation. Acidimetric titration showed the tetraethylammonium stock solution to be 0.673 M OH- and the LiOH stock solution to be 1.530 M OH-, and these stock solutions were maintained under argon. A 100 g / L Th stock solution in 0.1 M H N 0 3 was prepared. General Procedure. Solutions (20 mL) (series 1) of tetraethylammonium hydroxide above pH 12 were prepared by calculated dilution of the stock solution. To each solution was added 50 pL of the thorium stock solution (5 mg of Th). All solutions (20 mL) below p H 12 were prepared by adding 50 pL of the Th stock solution to water followed immediately by addition of tetraethylammonium hydroxide to the desired pH if above pH 9. For desired pH values below 9, the solutions were brought to pH 9-10 with tetraethylammonium hydroxide, causing immediate precipitation, and were brought back immediately to the desired pH with 0.1 M H N 0 3 . Six solutions (series 2) in the [OH-] range 0.28-1.38 M were prepared from the LiOH stock solution and 1.00 mL of Th stock solution (100 mg of Th). At this Th concentration, the Th/Si ratio in these samples varied from 235 to 935, and thus it is strongly felt that Si impurity cannot affect the results. An additional 17 solutions (series 3) were prepared in the pH range 3.5-10 in which 0.625 mL of the Th stock solution (62.5 mg of Th) was added to water that was then made basic with carbonate-free NaOH. The precipitated hydrous T h o 2 was washed with water and made up to 17 mL of 0.1 M NaC10, and the desired p H with hydrochloric acid or NaOH solution. Solution makeup was done under an Arcover flow or in an inert-atmosphere box. Solutions were immediately sealed; series 1 and 2 were shaken for 25-27 days and series 3 was shaken for 7 days at 25 O C . Solutions were centrifuged in the sealed containers for from 1 to 4 h at 1000 G. This removed any detectable Tyndall effect from most of the high-pH samples but not from the lower pH solutions. Ten-milliliter portions of the centrifuged series 1 and 2 solutions were then filtered at 60-psi pressure through 13-mm-diameter Nuclepore Ultrafilters, Type F, No. 1F7257, having mol wt 20 000 cutoff for globular proteins (- 2nm pore size). Series 3 solutions were filtered through Amicon C F 25 conical centrifuge filters having mol wt 25 000 (for globular proteins) cutoff. Filtration times varied from 20 min to 3 h in proportion to the extent of the Tyndall effect observable after centrifugation. The pH values of the unfiltered portions of the solutions were determined at the time of sampling for those solutions having pH 512, while those above 12 were calculated from the measured OH- concentration. Analyses. Thorium analyses on the filtered series 1 and 2 solutions were done by neutron activation. Samples, along with Th standards and a USGS GSP-1 standard, were irradiated at -6 X 10l2neutrons cm-2 s-' for 14 h. Samples were counted on a normal Ge(Li) counting system and on a coincidence and anticoincidence Ge(Li)-NaI(T1) system. Approximately 4 weeks' delay between irradiation and counting was necessary to allow for decay of 24Na ( t I l 2= 15 h) and 82Br = 36 h). Bromine is a possible expected impurity in tetraethylammonium hy-
0 1987 American Chemical Society
Notes droxide on the basis of its method of manufacture and indeed could be detected turbidimetrically by Ag' precipitation. After decay of these impurities, Compton scattering from (rl/2 = 65 days) and 65Zn(rllz = 240 days) in the tetraethylammonium hydroxide solutions increased the background, preventing increase of the sensitivity for the Th activation product, (tll2 = 27 days), with the coincidenceanticoincidence system. The Sr was, of course, added to remove carbonate, but the Zn was an unsuspected impurity in the tetraethylammonium hydroxide. Thorium analyses on the filtered and acidified series 3 solutions and blanks to which no Th was added were done by inductively coupled plasma mass spectroscopy. A combination glass electrode calibrated with pH 4, 7, and 10 buffers was used for pH measurement.
Results The solubility of Th(1V) as a function of solution pH is shown in Figure 1. The values shown can be considered to be the solubility of amorphous (hydrous) T h o 2 since electron diffraction studied5 of aging of hydrous T h o z at 25 OC show no crystallinity a t the time period (25-27 days) or less of the present solubility study. Solubilities in tetraethylammonium hydroxide (series 1) in the pH region 10-14 are essentially all at or below the detection limit of about lo4 M. Measured values in LiOH solutions (series 2) in the pH range 13.4-14.2 were in the 10-'o-10-9M Th range. These values appear to represent true Th measurements in the neutron activation analysis, but whether they represent true solubilities is not certain since the points are limited in number and two values that were duplicates vary from each other by a factor of 6.6. Since solubilities in this region were approached from the supersaturated side, the values can all be considered upper limits. These results show that the soluble neutral species, Th(OH)40(aq),cannot contribute to total thorium solubility to an extent exceeding 2 X or at the absolute highest lo4 M. This contradicts the results of Nabivanets and K ~ d r i t s k a y a who , ~ reported a value of 4.8 X lo-' M for the pH-independent concentration of this species in equilibrium with amorphous Tho2. This difference probably indicates that their value is a background level in either their solids-liquid separation technique or their analytical technique. The data indicate that amphoteric species do not contribute to Th solubility of >lo4 M at pH 14,but whether the values measured in LiOH are due to amphoteric species at all is not certain. At pH values between 5.5 and 11, measured Th values were obtained for both series 1 and 3 samples, but all series 3 values in this range are within a factor of 1.6 of the average blank values so they can be considered to be at background. For the series 1 samples in this range, only one at pH 5.7is significantly above background, and it appears to be a flier. In this region, solubility was approached from undersaturation in most samples, and in the two lowest pH points of Figure 1, the volume of solids visibly decreased in the first hour of equilibration. In the lowest pH sample only a Tyndall effect remained. It is clear, however, that in the pH 3.6-4.7 region, the average slope is somewhat greater than 4. The fact that the slope is greater than 4 may be at least partially due to the presence of some polynuclear species, at least at the higher Th concentration, low-pH end of this region. Because of the difficulty of accurately determining slopes greater than 4 and because of possible complications due to a small pH-dependent anion-exchange capacity of hydrous Th02,16,17 no attempt is made here to interpret this slope. This steep slope over about 5.5 orders of magnitude in aqueous Th concentration does point to mononuclear hydrolyzed Th ions not being dominant solution species at least up to pH 4.5. Between pH 4.5 and 6 a lower slope (dl) is observed, but since it is based entirely on three points that are only a factor of 3 above background, we do not attempt to interpret it. Baes and Mesmer,' in reviewing several prior studies, have given formation constants for such polynuclear species. It should Beasley, M. L.; Milligan, W. 0. J . Electron Microsc. 1967, 16, 101. (16) Amphlett, C. B.; McDonald, L. A.; Redman, M. J. J . Inorg. Nucl. Chem. 1958, 6, 236. (17) Kraus, K. A,; Phillips, H. 0.;Carlson, T. A.; Johnson, J. S. Proc. Int. Conf. Peaceful Uses A t . Energy, 2nd 1958, 28, 3. (15) Prasad, R.;
Inorganic Chemistry, Vol. 26, No. 24, 1987 4141 be noted that other values have since been published2 but that the hydrolysis scheme should decrease the slope to less than 4. Using the values given by Baes and Mesmer,' we can show that polynuclear species will not be significant at pH 4.0 and 1.5 X M Th (the solubility at that pH for series 1 solutions). The solubility product of the amorphous T h o z in the present work can be determined from these numbers. Neglecting mononuclear complexes and activity coefficients would give an apparent log Ksp = -44.8 for the reaction Th02(am) + 2 H 2 0 s Th4+ + 40H(1) Series 1 data are used here since this same series (same hydrous oxide aging time etc.) extends to the very high pH range. It should be apparent from Figure 1 that essentially the same result is obtained from series 3 results in the pH 3.9-4.3range, where solubilities are low enough to minimize polynuclear species in solution but are high enough to provide good analytical accuracy. If the hydrolysis constants of Baes and Mesmerl for formation of Th(OH)3+and Th(OH)22+are correct, log Kspwould become more negative by 0.7 and the activity coefficient correction for our ionic strength of -0.005 would make long Kspmore negative by an additional 0.4,giving a value of log K!p = -45.9. There is some reason, in addition to the >4 slope in Figure 1, to question whether the use of the Baes and Mesmer values of the two formation constants, particularly the second, is completely valid. As we show below, the values for Th(OH):(aq) and Th(OH)5- are very much smaller than the proposed values. If one does accept the unproven concept proposed by Baes and Mesmer' as well as by Barnum4 that there is a continuous linear progression in differences in the logarithms of formation constants, the value reported for Th(OH)?+ would appear too high. Nevertheless, this should not produce an error of more than about 0.5 in log Ksp. (It should also be noted that even if the hydrolysis scheme of Brown et aL2 were used, our data would give log Ksp = -46.3.) For comparison, a value of log Ksp= -46.6 was calculated by Baes and Mesmer' for the amorphous T h o 2 in the work of Nabivanets and K ~ d r i t s k a y a . ~ The log (solubility product) of fully crystalline Tho2, eq 1 with Th02(c) replacing Th02(am), calculatedIs from accepted thermochemical data5-' for ThOz(c) (AGf = -1 168.8kJ mol-') and Th4+ (AGf = -704.6 kJ mol-'), is log Ksp = -54.14. This value is considerably lower than the value of -49.4 reported by Baes and Mesmerl to have been calculated from thermochemical data, and the source of their data is unclear. It also is considerably lower than the reported value of -49.7 based on solubility measurements of Baes et al.,I9 and it appears very unlikely on the basis of the electron diffraction data of Pmsad et al.15 that their T h o 2 approached from supersaturation could be completely and fully crystalline in the time frame of their experiments. It must be emphasized that measured solubilities will always be controlled by any trace of less than fully crystalline material present. On the basis of the difference between Kspfor Th02(c) from thermochemical data and the value for our Th02(am), we obtain a difference of 47.1 kJ mol-' between the free energies of formation of these oxides. This agrees reasonably well with a free energy difference of 41.4kJ mol-' between our value for P U O , ( ~ ~ ) ' * ~ ~ ~ and the thermochemical value for P U O , ( C ) and ~ ~ ~a free energy difference of 52.5 kJ mol-' between the value for NpO2(am)I4 and the thermochemical value for Np02(c) The data of Figure 1 give log K,, d -9.0 for the reaction Th02(am) 2 H 2 0 OH- G Th(OH)5(2) .596
+
+
From the free energy of formation (-1121.7 kJ mol-') of ThO,(am) determined from the Kspdata above, a value of AGf 2 -1702.1 kJ mol-' for Th(OH)5- is obtained. This value, along with the free energy of formation of Th4+ cited above, leads to log p5* d -33.1 for the reaction Th4+ + 5H20 + Th(OH)5- 5H+ (3)
+
J. L. Radiochim. Acta, in press. (19) Baes, C. F.; Meyer, N. J.; Roberts, C. E. Inorg. Chem. 1965, 4 , 518. (20) Rai, D.Radiochim. Acta 1984, 35, 97. (18) Rai, D.; Swanson, J. L.; Ryan,
4142 Inorganic Chemistry, Vol. 26, No. 24, 1987
Additions and Corrections
Similarly, from the data of Figure 1, a value of AGf 3 -1540.7 kJ mol-I for Th(OH),O(aq) and a value of log p4* G -19.7 are obtained for the reaction Th4+
+ 4 H 2 0 F! Th(OH),O(aq) + 4H'
(4)
These values result in calculated solubilities of Th02(c) of G1.2 lo-'* M as the pH-independent species, Th(OH),O(aq), and C6 X lo-'* M at pH 14 as the Th(OH)< ion. In contrast, Baes and Mesmer' propose a pH-independent minimum Th02(c) solubility as Th(OH),O(aq) of 2.5 X M, and Barnum's4 data for Th(OH),- yield a Th02(c) solubility of 0.4 M at pH 14. This error of a factor of 36.6 X 10l6 in solubility is much greater than Barnum's claimed accuracy in his formation constants of about a factor of 3.4 Such inaccuracies put in extreme doubt the validity of Barnum's4 and Baes and Mesmer's' proposed linear or nearlinear extrapolation of differences in logarithms of hydrolysis constants or free energies of formation of mononuclear hydrolytic species. We have now placed limits on the logarithms of formation constants (log &*) of Th(OH)5-, U(OH)5-, and N P ( O H ) ~ -of G33.1, C-22.7,I3 and C-24.7,I4 respectively. The formation constant of Pu(OH),- has been reported in several recent publ i c a t i o n ~ , ' ~ ~ ~and ~ - 'a' ~value ~ ' of log &* = -15.0 is usually give n . ' ~ ~ ,Such ~ - ~a~value, when combined with AGf for Pu4+ of -481.6 kJ mol-' 22 and a value of log Ksp= -56.85 for amorphous X
(21) Kim, J. I.; Lierse, C.; Baumgartner, F. ACS Symp. Series, 1983, No. 216, 317. (22) Fuger, J.; Oetting, F. L. The Chemical Thermodynamics of Actinide Elements and Compunds-II. The Actinide Aqueous Ions; lnternational Atomic Energy Agency: Vienna, 1976.
Pu02,18,20 would lead to a solubility of Pu02(am) of >0.01 M at pH 14. Kim et al.2' have actually shown calculated solubilities at pH 14 considerably in excess of 1 M Pu(1V). Such values are preposterous in terms of the long known fact that Pu(IV) is quantitatively precipitated by excess alkali, and such formation constants should not be published. It is not possible to extrapolate our limit values for Th, U, and Np to Pu as a function of ionic radii precisely because they are only upper limits. It is, however, reasonable, on the basis of known chemical behavior in other be systems, to expect that the log p5* value for P U ( O H ) ~would no more than about 2 log units more positive than that for Np(OH),-. This would place log &* 6 -22.7 for Pu(OH),- and M would yield a Pu02(am) solubility as Pu(IV) of 6 2 . 3 X at pH 14. In view of the oxidation-state-control problem with Pu(1V) under alkaline conditions, such a value would be difficult to reliably verify. In fact, it is very probable that a pH region exists in which Pu(IV) is never the dominant solution oxidation state regardless of redox potential. This region may extend as low as pH 4 and as high as pH 14. On the basis of this work, we feel that no conclusive evidence of any amphoteric behavior of any tetravalent actinide in aqueous solutions has been published. In conclusion, a value of log Ksp = -45.9 A 0.5 for Th02.xH20 was obtained from these data. The results set upper limits for log &* 6 -33.1 for the reaction Th4+ + 5 H 2 0 s Th(OH)*- + 5H' and for log p4* G -19.7 for the reaction Th4+ 4 H 2 0 s Th(OH),O(aq) + 4H'. Acknowledgment. This research was conducted for the U S . Department of Energy under Contract DE-AC06-76RLO 1830. We thank Dr. J. C. Laul for neutron activation analyses.
+
Registry No. T h o z , 1314-20-1; LiOH, 1310-65-2; tetraethylammonium hydroxide, 77-98-5.
Additions and Corrections 1987, Volume 26 Randolph P. Thu"el,*
Francois Lefoulon, and James D. Korp: Polyaza Cavity-Shaped Molecules. 12. Ruthenium(I1) Complexes of 3,3'-Annelated 2,2'-Bipyridine: Synthesis, Properties,and Structure. Pages 2370-2376. The correct space group for complex 4d is Pi and not P1 as given on p 2370, abstract, line 7; p 2374, paragraph 3, line 1; and p 2375, Table VII, line 3. The Laue symmetry for 4d is and not 1 as given on p 2375, paragraph 4, line 6.-Randolph P. Thummel
